Würzburg 2018 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

AKPIK: Arbeitskreis Physik, moderne Informationstechnologie und Künstliche Intelligenz

AKPIK 1: Arbeitskreis Physik, IT & KI (AKPIK) I

AKPIK 1.6: Vortrag

Montag, 19. März 2018, 18:00–18:15, 70 - HS 00.107

Deep Neural Networks for IceCube Online Classification — •Jan Spinne, Philipp Hoffmann, and Felix Neubürger for the IceCube collaboration — TU Dortmund

IceCube is a neutrino detector located at the geographic South Pole, which detects Cherenkov light emitted by charged secondary particles passing through the ice. A key challenge for the real-time analysis framework in IceCube is the classification of events based on their corresponding topology. For example, topologies can include cascade-like events for electron neutrino interactions and track-like events of charged-current muon neutrino interactions. The detector's high event rate and limited computational resources available on site, complicate the classification task. Real-time analyses should therefore be fast and effective.

This talk aims to motivate and show the usage of deep neural networks, which are predestined due to its fast model application and potentially high accuracy.